Empirical findings show that access to adequate and sufficient food in developing countries is unstable, suggesting that whether a household or individual is food secure at any point in time is best thought of in a dynamic sense. The more widely used food security analysis methods mainly consider current access to food, failing to provide policy makers with forward-looking information. While drawing on household survey data, the vulnerability analysis model presented here provides estimates of the probability that a given household will lose or gain access to sufficient food in the near future. We propose a model of vulnerability analysis that can improve policy design and targeting. We test the model with data from a survey of Nicaraguan households.